I am experience a clone controlled replication slow performance, the bandwidth is 100 MB between two sites, and replication performed around 8~10 MB, no any throttling setup, "low bandwidth optimization" is not enabled, but it automatically showed in the log report (under the column, it shows 1), suspect the "low bandwidth optimization" is enabled by NW CCR and it automatically set throttling.
Any one can shed light?
Hm... are you sure you have 100MB? Or is that 100Mbps which would give you what you see indeed? If you were referring to more, I would start by checking cables (for example using wrong CAT can give you what you see).
in case the DD's have multiple NICs, are you sure that CCR replication uses the intended interface? Only way to specifically control which interfaces to use for replication, is by specifying the DD replication interface to use in the hosts file of the other DD and vice versa.
So the name of target DD as defined within NW under the ddboost device "device access information", needs to be put into the hosts file of the source DD with the ip address of the target DD. And on the source DD, the replication ip address of the target DD needs to be defined with the target DD name as stated in NW device acces information of the targets ddboost device.
I can't think of anything that would cause some kind of low bandwidth to be enforced, if you havent set neither lowbandwidth optimization not any throttling. unless the reported bandwidth is not as high as stated (for instance 100Mb instead of 100MB).
One thing to test could be iperf testing (in the so-called SE mode of a DD), running one of the DD's in iperf server mode and the other one as sending client. Then you could find out the the communication speed between the 2 replication interfaces without ddboost.
ble1
4 Operator
•
14.4K Posts
0
December 2nd, 2013 12:00
Hm... are you sure you have 100MB? Or is that 100Mbps which would give you what you see indeed? If you were referring to more, I would start by checking cables (for example using wrong CAT can give you what you see).
bbeckers1
2 Intern
•
203 Posts
0
December 2nd, 2013 12:00
in case the DD's have multiple NICs, are you sure that CCR replication uses the intended interface? Only way to specifically control which interfaces to use for replication, is by specifying the DD replication interface to use in the hosts file of the other DD and vice versa.
So the name of target DD as defined within NW under the ddboost device "device access information", needs to be put into the hosts file of the source DD with the ip address of the target DD. And on the source DD, the replication ip address of the target DD needs to be defined with the target DD name as stated in NW device acces information of the targets ddboost device.
I can't think of anything that would cause some kind of low bandwidth to be enforced, if you havent set neither lowbandwidth optimization not any throttling. unless the reported bandwidth is not as high as stated (for instance 100Mb instead of 100MB).
One thing to test could be iperf testing (in the so-called SE mode of a DD), running one of the DD's in iperf server mode and the other one as sending client. Then you could find out the the communication speed between the 2 replication interfaces without ddboost.